Farmers’ perceptions, adoption and impacts of integrated water management technology under changing climate

Figures & data

Figure 1. Influencing pathways of perceptions, adoption and impacts of integrated water management technology (IWMT).

Table 1. Variable definition and summary statistics.

Variables	Definitions	Mean (SD)
Outcome variables
Banana yields	Yields of banana production (1000 kg/mu)^a	3.18 (2.83)
Gross revenue	Gross revenue of banana production (1000 yuan/mu)^b	7.10 (7.21)
Net returns	Gross revenue minus investment costs of banana production (1000 yuan/mu)	3.93 (7.24)
Irrigation frequency	Frequency of irrigation during banana production (×100)	0.34 (0.45)
Irrigation expenditure	Expenditure on irrigation (100 yuan/mu)	0.53 (1.40)
Treatment variable
IWMT adoption	1 = Adopters of integrated water management technology (IWMT); 0 = otherwise	0.52 (0.50)
Control variables
Age	Age of household head (years)	48.20 (9.90)
Gender	1 = Male; 0 = otherwise	0.83 (0.37)
Education	Educational level (years)	8.08 (3.12)
Cadre experience	1 = Is or has been a village cadre; 0 = otherwise	0.23 (0.42)
Family size	Number of family members (persons)	5.98 (2.37)
Land size	Total orchard size for banana production (mu)	29.33 (72.93)
Drought perception	Perceived drought frequency compared with five years ago: 1 = More frequent; 2 = Almost; 3 = Less frequent	2.10 (0.61)
Clay soil	1 = Orchard mainly has clay soil; 0 = otherwise	0.50 (0.50)
Loam soil	1 = Orchard mainly has loam soil; 0 = otherwise	0.29 (0.45)
Sandy soil	1 = Orchard mainly has sandy soil; 0 = otherwise	0.21 (0.41)
Distance to market	Distance to the nearest input market (km)	5.72 (8.46)
Distance to credit sources	Distance to the nearest formal (e.g. bank and financial institutions) and informal (e.g. relatives or financial groups) credit source (km)	5.89 (8.33)
Hainan	1 = Located in Hainan province; 0 = otherwise	0.37 (0.48)
Yunnan	1 = Located in Yunnan province; 0 = otherwise	0.29 (0.46)
Guangdong	1 = Located in Guangdong province; 0 = otherwise	0.34 (0.47)
Observations		626

Note: ^a1 mu = 1/15 ha.

^bYuan is Chinese currency (US$1 = 6.90 yuan in 2019).

SD, standard deviation.

Table 2. Mean differences of variables between integrated water management technology (IWMT) adopters and non-adopters.

Variables	IWMT adopters	Non-adopters	Mean differences
Outcome variables
Banana yields	3.60 (3.34)	2.73 (2.06)	0.87***
Gross revenue	7.66 (6.78)	6.49 (7.60)	1.17**
Net returns	4.19 (6.96)	3.64 (7.53)	0.55
Irrigation frequency	0.49 (0.53)	0.18 (0.24)	0.31***
Irrigation expenditure	0.48 (1.23)	0.57 (1.57)	−0.08
Control variables
Age	47.36 (10.12)	49.10 (9.59)	−1.74**
Gender	0.81 (0.39)	0.86 (0.35)	−0.05*
Education	8.15 (3.16)	8.01 (3.08)	0.14
Cadre experience	0.22 (0.41)	0.24 (0.43)	−0.02
Family size	6.07 (2.43)	5.87 (2.29)	0.20
Land size	30.42 (75.11)	28.16 (70.63)	2.26
Drought perception	2.15 (0.69)	2.04 (0.50)	0.12**
Clay soil	0.60 (0.49)	0.40 (0.49)	0.20***
Loam soil	0.20 (0.41)	0.37 (0.49)	−0.17***
Sandy soil	0.20 (0.40)	0.23 (0.42)	−0.03
Distance to market	3.41 (3.21)	8.17 (11.19)	−4.77***
Distance to credit sources	4.51 (4.22)	7.37 (10.97)	−2.86***
Hainan	0.51 (0.50)	0.21 (0.41)	0.31***
Yunnan	0.10 (0.29)	0.51 (0.50)	−0.42***
Guangdong	0.39 (0.49)	0.28 (0.45)	0.11***
Observations	323	303

Note: ***p < 0.01, **p < 0.05 and *p < 0.10.

Standard deviation (SD) is presented in parentheses.

Figure 2. Farmers’ perceptions of fertilizer and water-saving levels through integrated water management technology (IWMT) adoption: (a) perceptions of fertilizer-saving levels; and (b) perceptions of water-saving levels.

Figure 3. Farmers’ perceptions of the integrated water management technology (IWMT) adoption with respect to technology similarity, operability, economic benefit and others’ attitudes: (a) technology similarity; (b) operability; (c) economic benefit; and (d) others’ attitudes.

Table 3. Pearson correlation coefficients between perceptions variables and integrated water management technology (IWMT) adoption.

Variables	IWMT adoption	p-value
Workload reduction	0.243***	0.000
Fertilizer reduction	0.255***	0.000
Water reduction	0.103**	0.010
Water scarcity	0.014	0.727
Technology similarity	0.029	0.465
Operability	0.514***	0.000
Economic benefit	0.250***	0.000
Others’ attitude	0.589***	0.000

Note: ***p < 0.01 and **p < 0.05.

Table 4. Average treatment effects of integrated water management technology (IWMT) adoption on outcome variables: propensity score matching (PSM) estimator.

	Mean outcomes
Outcome variables	Actual	Counterfactual	ATT	% Change
Nearest-neighbour matching (NNM) (1–3 matching)
Banana yields	3.604	2.984	0.620 (0.252)**	20.78
Gross revenue	7.648	6.247	1.401 (0.524)***	22.43
Net returns	4.195	3.141	1.054 (0.563)*	33.56
Irrigation frequency	0.489	0.262	0.226 (0.045)***	86.26
Irrigation expenditure	0.490	0.365	0.126 (0.182)	34.52
Kernel matching
Banana yields	3.604	3.019	0.585 (0.218)***	19.38
Gross revenue	7.648	6.190	1.457 (0.522)***	23.54
Net returns	4.195	3.140	1.055 (0.523)**	33.60
Irrigation frequency	0.489	0.270	0.218 (0.041)***	80.74
Irrigation expenditure	0.490	0.412	0.078 (0.187)	18.93

Note: ***p < 0.01, **p < 0.05 and *p < 0.10.

ATT, average treatment effects on the treated.

100 bootstrapped standard errors are presented in parentheses.

Table 5. Average treatment effects by gender of household head and land size: nearest-neighbour matching (NNM) (1–3 matching).

		Mean outcomes
Groups		Actual	Counterfactual	ATT	% Change
By gender of household head
Male	Banana yields	3.319	3.054	0.265 (0.215)	8.68
	Gross revenue	7.288	6.055	1.233 (0.563)**	20.36
	Net returns	3.862	3.111	0.751 (0.701)	24.14
	Irrigation frequency	0.483	0.310	0.177 (0.041)***	57.10
	Irrigation expenditure	0.522	0.394	0.128 (0.192)	32.49
Female	Banana yields	4.744	3.097	1.645 (0.873)*	53.12
	Gross revenue	9.065	5.809	3.256 (1.744)*	56.05
	Net returns	5.523	2.309	3.214 (1.855)*	139.19
	Irrigation frequency	0.517	0.290	0.227 (0.122)*	78.28
	Irrigation expenditure	0.368	0.536	−0.168 (0.434)	−31.34
By land size
Tertile 1 (0.5–5 mu)	Banana yields	4.396	3.170	1.226 (0.493)**	38.68
	Gross revenue	8.389	5.886	2.504 (1.291)*	42.54
	Net returns	5.004	2.673	2.331 (1.591)	87.21
	Irrigation frequency	0.462	0.207	0.255 (0.067)***	123.19
	Irrigation expenditure	0.543	0.034	0.509 (0.157)***	1497.06
Tertile 2 (5.5–15 mu)	Banana yields	3.335	3.021	0.315(0.314)	10.43
	Gross revenue	6.956	5.714	1.242 (0.788)	21.74
	Net returns	3.668	2.968	0.701 (0.926)	23.62
	Irrigation frequency	0.552	0.320	0.232 (0.072)***	72.50
	Irrigation expenditure	0.443	0.902	−0.459 (0.534)	−50.89
Tertile 3 (16–1000 mu)	Banana yields	3.087	2.338	0.749 (0.306)**	32.04
	Gross revenue	7.992	5.962	2.029 (0.961)**	34.03
	Net returns	4.286	2.484	1.802 (1.281)	72.54
	Irrigation frequency	0.452	0.304	0.149 (0.153)	49.01
	Irrigation expenditure	0.437	0.145	0.022 (0.591)	15.17

Note: ***p < 0.01, **p < 0.05 and *p < 0.10.

ATT, average treatment effects on the treated.

100 bootstrapped standard errors are presented in parentheses.

Figure A1. Distributions and common support for propensity score matching (PSM).

Table A1. Variable definition and summary statistics of perception variables.

Perception variables	Definitions	Mean (SD)
Workload reduction	Do you perceive that IWMT adoption reduces the workload allocated for irrigating farmland and applying fertilizers? (1 = Yes, 0 = No)	0.79 (0.41)
Fertilizer reduction	Do you perceive that IWMT adoption helps reduce fertilizers? (1 = Yes, 0 = No)	0.68 (0.47)
Water reduction	Do you perceive that IWMT adoption helps reduce water used for farmland irrigation? (1 = Yes, 0 = No)	0.55 (0.50)
Water scarcity	Did you experience irrigation water scarcity during the 2017/18 banana production year? (1 = Yes, 0 = No)	0.39 (0.49)
Technology similarity	To what extent do you perceive there exists a difference between IWMT and furrow irrigation technology (1 = Totally different; 2 = Somehow different; 3 = Neutral; 4 = Almost the same; 5 = Totally the same)	2.63 (1.00)
Operability	To what extent do you perceive that IWMT is easier to operate in practice than furrow irrigation? (1 = Very difficult; 2 = Difficult; 3 = The same; 4 = Easy; 5 = Very easy)	3.54 (1.04)
Economic benefit	To what extent do you perceive that IWMT adoption generates higher economic benefits than furrow irrigation technology? (1 = Much lower; 2 = Lower; 3 = The same; 4 = Higher; 5 = Much higher)	3.75 (0.79)
Others’ attitude	To what extent do you perceive that other farmers’ attitudes towards IWMT adoption are supportive? (1 = Very opposite; 2 = Opposite; 3 = Neutral; 4 = Supportive; 5 = Very supportive)	3.70 (0.90)
Observations		626

Note: SD, standard deviation.

Table A2. Matching quality test: balancing property.

	Before matching	After matching
		Nearest-neighbour matching (NNM) (1–3 matching)	Kernel matching
Pseudo-R^2	0.220	0.013	0.008
Mean bias	31.7	6.1	5.1
LR χ^2	190.79***	11.26	6.95
p-value	0.000	0.589	0.905

Note: ***p < 0.01.

Table A3. Propensity score matching (PSM) and covariate balance: nearest-neighbour matching (NNM) (1–3 matching).

Variable	Matched	Treated	Control	% bias	% reduction in bias
Age	Unmatched	47.36	49.10	−17.6%
	Matched	47.53	48.17	−6.4%	63.5
Gender	Unmatched	0.81	0.86	−13.4%
	Matched	0.81	0.82	−2.5%	81.2
Education	Unmatched	8.15	8.01	4.6%
	Matched	8.12	8.16	−1.3%	71.3
Cadre experience	Unmatched	0.22	0.24	−5.8%
	Matched	0.22	0.25	−6.7%	−16.6
Family size	Unmatched	6.07	5.87	8.6%
	Matched	6.07	6.43	−15.2%	−76.5*
Land size	Unmatched	30.42	28.16	3.1%
	Matched	27.30	30.90	−4.9%	−59.2
Drought perception	Unmatched	2.15	2.04	19.2%
	Matched	2.16	2.11	8.5%	55.6
Clay soil	Unmatched	0.60	0.40	41.1%
	Matched	0.60	0.64	−7.5%	81.8%
Loam soil	Unmatched	0.21	0.38	−37.7%
	Matched	0.21	0.18	6.8%	82.0
Distance to market	Unmatched	3.41	8.17	−57.9%
	Matched	3.43	3.01	5.1%	91.2
Distance to credit sources	Unmatched	4.51	7.37	−34.5%
	Matched	4.52	3.97	6.7%	80.6
Hainan	Unmatched	0.51	0.21	67.1%
	Matched	0.51	0.48	6.9%	89.8
Yunnan	Unmatched	0.09	0.51	−101.4%
	Matched	0.09	0.10	−1.3%	98.7

Note: *p < 0.10.

Table A4. Rosenbaum bounds for sensitivity analysis of IWMT adoption on outcome variables.

Outcome variables	Gamma	t-hat+	t-hat–	CI+	CI–
Banana yields	1	3.063	3.063	2.875	3.260
	1.2	2.925	3.208	2.742	3.420
	1.4	2.813	3.336	2.625	3.567
	1.6	2.716	3.458	2.533	3.700
	1.8	2.625	3.550	2.458	3.825
	2	2.563	3.653	2.378	3.940
Gross revenue	1	6.610	6.610	6.148	7.145
	1.2	6.273	7.000	5.800	7.530
	1.4	6.000	7.320	5.522	7.900
	1.6	5.750	7.613	5.313	8.220
	1.8	5.545	7.875	5.114	8.520
	2	5.387	8.110	4.943	8.800
Net returns	1	3.239	3.239	2.764	3.736
	1.2	2.893	3.595	2.425	4.125
	1.4	2.610	3.906	2.147	4.483
	1.6	2.372	4.200	1.908	4.800
	1.8	2.167	4.457	1.709	5.100
	2	1.986	4.699	1.525	5.375
Irrigation frequency	1	0.415	0.415	0.380	0.450
	1.2	0.390	0.450	0.350	0.490
	1.4	0.365	0.475	0.330	0.510
	1.6	0.350	0.500	0.310	0.540
	1.8	0.330	0.510	0.300	0.550
	2	0.320	0.540	0.280	0.595
Irrigation expenditure	1	<0.001	<0.001	<0.001	0.125
	1.2	<0.001	0.100	<0.001	0.250
	1.4	<0.001	0.182	<0.001	0.375
	1.6	<0.001	0.273	<0.001	0.435
	1.8	<0.001	0.375	<0.001	0.500
	2	<0.001	0.417	<0.001	0.571

Note: N = 622 matched pairs.

Gamma is log odds of differential assignment due to unobserved factors.

t-hat+ is the upper bound Hodges–Lehmann point estimate;

t-hat– is the lower bound Hodges–Lehmann point estimate;

CI+ is the upper bound confidence interval (a = 0.95); and

CI– is the lower bound confidence interval (a = 0.95).

Table A5. Average treatment effects of IWMT adoption on outcome variables: IPWRA estimator.

	Mean outcomes
Outcome variables	Actual	Counterfactual	ATT	% Change
Banana yields	3.598	3.009	0.589 (0.217)***	19.57%
Gross revenue	7.660	6.136	1.524 (0.467)***	24.84%
Net returns	4.194	3.096	1.098 (0.511)**	35.47%
Irrigation frequency	0.491	0.277	0.214 (0.037)***	77.26%
Irrigation expenditure	0.484	0.430	0.054 (0.160)	12.56%

Note: ***p < 0.01; and **p < 0.05.

ATT, average treatment effects on the treated.

Robust standard errors are presented in parentheses.